Student Learning Analytics Based Prediction Framework

Abstract

Student performance at any level of academic period depends on monitoring parameters. Selection of affecting parameter depends on student various parameters like social, effort, teaching techniques, etc. This work focus on finding the student performance patterns under three observations Good, Average and Poor.Based on academic patterns features were select which highly affect the student performance by genetic algorithm. Dynamic adoption of genetic algorithm gives a feasible feature set to predict the student performance. Feature set were further process to train the multi layerneural network that act as flexible framework for predicting the student performance. Maharashtra Talent Search Examination (MTSE) dataset was used for experimental work, having 0.1 million participant in a year. Results shows that proposed Student Learning Analytics Based Prediction Framework (SLAPF) has improved comparison parameters from existing student performance prediction models